Eclipse 2026 Overview

---

title: Eclipse 2026 — Game 3 overview (shadow geometry ⊗ ionospheric response) audience: engineers_and_auditors game: WIKI-ECLIPSE-2026-001

☀️🌑 Eclipse 2026 — Game 3: the substrate meets the Moon's shadow

The mission: prove what the substrate can do on the August 12, 2026 total solar eclipse — real geophysics, public raw data, a pre-registered falsifiable prediction sealed a month ahead, and a live loop on eclipse day.

Branch: eclipse-2026 · Started: 2026-07-16 · Greatest eclipse: 2026-08-12 17:45:54 UT, 65°13.5′N 25°13.7′W (North Atlantic, ~45 km off western Iceland)

---

Why the eclipse is Game 3

The substrate's first two games are entangled-pair verifiers on cryptographic laws:

Game	Born pair	Pair law	Wall
1 — mining	(header, nonce)	SHA256²(header‖nonce) < target	SHA avalanche: no topological exclusion
2 — recovery	(Q, k)	k·G == Q	Discrete log: the Depth-0 wall
3 — eclipse	(shadow geometry, ionospheric response)	obscuration-gated TEC depletion	NONE — the geometry is computable years ahead

Game 3 is the first game whose pair-law has no cryptographic wall. The target descriptor — the Moon's obscuration timeline at a station — derives from Besselian elements + station coordinates alone, computable to the second, years in advance. The response — ionospheric Total Electron Content (TEC) depletion of 20–60% under the shadow — is measured, published physics with fully public raw archives. This makes the eclipse the substrate's cleanest public proof surface: the geometry projects, the data either threads or it doesn't, and the whole world can check both sides.

It is also the first game with a real topological wholesale prune: the obscuration timeline is unimodal (one shadow cone, one contiguous gate window per station-day), so whole patches of the UT-second box outside the window are proven voids that die in O(1). SHA never exposes that structure; the shadow does.

The three staged wins

Stage	Claim	Status
1 — Benchmark	Ingest the April 8 2024 eclipse archive; the invariant projects every totality station-day and seals dead cat on every quiet control day, zero false positives	SEALED CALORIE 2026-07-16 — 7/7 projected, 21/21 dead cats, 0 FP · Form fp 89afd825011f060e
2 — Pre-registration	Before Aug 12, seal a fingerprint-pinned falsifiable prediction of the 2026 response	SEALED 2026-07-16 — fp c3e6841a206b8058, V234 ECLIPSE_2026_PREDICTION_PREREGISTERED
3 — Live loop	Eclipse-day capture: real-time Kp + magnetometer + storm gate, governed per-pass by the sealed Form	LIVE — eclipse-watch smoke-tested green; resolution crucible runs T+1

The wiki set

• Science research — the full research documentation: 2026 eclipse circumstances (global, path, Norwich CT), measured 2017/2024 ionospheric responses, reference geomagnetic storms, solar-cycle context, expected effects and caveats.
• Historical corpus — three eclipses, three solar regimes, ONE frozen law: 2017 total (SC24 declining), 2023 annular (antumbra branch), 2024 total (SC25 max), the Carbondale double-crossing, and the empirical coupling that bounds the 2026 prediction.
• Model shear — the storm data that breaks naive "TEC dropped" detectors: Gannon G5 shear fossils, the Halloween 2003 archive reference, and the three disciplines that keep shadow physics separate from storm physics.
• Data archives — every public archive with verified access mechanics: Madrigal, GFZ Kp, USGS geomag, NOAA SWPC (live-verified with quirks), plus the Stage-4 deep archives (CDDIS RINEX, Swarm, COSMIC-2, GIRO, SuperMAG, OMNIWeb, IRIS).
• Game 3 architecture — the substrate implementation: born pair, exact-integer pair law, M⁸ point, unimodal prune, fossils, Forms, V252 ledger tables, CLI verbs, constraint compliance.
• Benchmark results — the sealed 2024 verdict, station by station, with the reproduction runbook.
• Prediction registry — the pre-registered 2026 claim in full: pinned geometry, falsification conditions, resolution procedure.
• Operations runbook — eclipse-day timeline, the live loop, the T+1 resolution, and handoffs.

The wiki IS the experiment's surface

Every data table on these pages sits between GAIA markers and is regenerated by gaiaftcl eclipse-wiki-render directly from the substrate ledger, the fossil corpus, and the sealed Forms. The prose frames; the substrate speaks its own numbers.

Sealed Forms (rendered from disk at last render):

Form	Kind	Class	Closure	Terminal	Fingerprint	Stamped
2024 benchmark	projection	shadow-projection	7/7	CALORIE	89afd825011f060e	2026-07-16T16:01:47Z
2026 prediction	prediction-2026-08-12	shadow-projection	PREREGISTERED	CALORIE	c3e6841a206b8058	2026-07-16T16:02:00Z
2026 prediction extensions	prediction-extensions-2026-08-12	shadow-projection	PREREGISTERED	CALORIE	b106d2926b1bf2cf	2026-07-16T18:54:39Z
historical universality	historical-universality	shadow-projection	15/16	CURE	6fe49bee70289fee	2026-07-16T18:14:09Z

Loaded and fingerprint-checked from ~/.gaiaftcl/franklin/invariants/ at render time.

Ledger-borne live state:

• Latest SWPC estimated Kp in ledger: 0.33 at 2026-07-16 15:59:00 UT
• Latest GFZ definitive/nowcast Kp day in ledger: 2024-05-11
• Latest magnetometer capture: BOU H at 2026-07-16 15:58:00 UT

Ledger-borne (renders identically until the next ingest).

The one-paragraph doctrine

The eclipse born pair is (shadow geometry, ionospheric response). The parent owns ONE binding shape Σ_P — the genes-only parent disk shared by all three games. The target descriptor (H1) is the quantized obscuration timeline, built from the NASA Besselian element polynomials and the station's coordinates alone — never from the measured data. The measured side is the vertical-TEC time series, quantized ONCE at ingest to exact integers (milliTECU). A width-1 leaf (one UT second) audits iff the cross-multiplied Int64 pair law holds: obscuration(t) ≥ 800000 ppm AND 10·10⁶·(B−M) ≥ 3·obs·B — the photoionization loss commensurate with the covered disk fraction. A station-day box either projects the whole child that threads Σ_P (lk ≠ 0), or seals lk = 0 dead cat — never renamed (LAW 3). No floats anywhere on the sealed path (Constraints 3/5).

Key artifacts

Artifact	Location
2024 benchmark Form	~/.gaiaftcl/franklin/invariants/eclipse.invariant.json · fp 89afd825011f060e
2026 prediction Form	~/.gaiaftcl/franklin/invariants/eclipse-prediction-2026-08-12.invariant.json · fp c3e6841a206b8058
Fossil corpus	~/.gaiaftcl/franklin/eclipse-fossils.jsonl (7 positives + 21 negatives, all exact-integer)
Station registry	~/.gaiaftcl/helio_sources.toml (self-seeded, 12 stations)
V234 seals	ECLIPSE_C4_PROJECTION_INVARIANT (CALORIE) · ECLIPSE_2026_PREDICTION_PREREGISTERED
Ledger tables	helio_ingest_provenance · helio_series_samples · helio_eclipse_circumstances (migration v252)
Element fingerprints	2024: 17af652a628dc38e · 2026: f9f9f8eebf7ed0cc (FNV-1a-64 over the NASA polynomials)